As we want to centralized cgroup access we should stop killing the user
sessions directly from the systemd-user-sessions service. Instead, rely
on PID 1 doing this by adding the right ordering dependencies to the
session scope units.
When PID 1 reloads the units logind/machined will see UnitRemoved
signals for all units. Instead of trusting these immediately, let's
check the actual unit state before considering a unit gone, so that
reloading PID 1 is not mistaken as the end of all sessions.
Since we'll unload all units/job during a reload, and then readd them it
is really useful for clients to be aware of this phase hence sent a
signal out before and after. This signal is called "Reloading" (despite
the fact that it is also sent out during reexecution, which we consider
a special case in this context) and has one boolean parameter which is
true for the signal sent before the reload, and false for the signal
after the reload. The UnitRemoved/JobRremoved and UnitNew/JobNew due to
the reloading are guranteed to be between the pair of Reloading
messages.
When we issue a reexecution request via the private socket we need to
expect a "Disconnected" in addition to "NoReply" when the connection is
terminated.
When a scope unit is created due to deserialization rather than client
request don't enforce that the PIDs set must be non-empty, since the
cgroup is already populated.
While we are reloading we shouldn't suppress adding units to the bus
queue when there are no subscribers, simply because we might not have
deserialized the subscribers list yet. Hence, during reloading always
assume we have subscribers.
Some units set KillMode=none to survive the initrd→rootfs transition. We
cannot remove their cgroups, but that shouldn't really be considered an
issue, so let's downgrade the error message.
Almost everyone wants kmod support, so don't fail silently if the libs are
out-of-date.
kmod can still be explicitly disabled and if it is not found at all, we still
default to disabling it.
As of kmod v14, it is possible to export the static node information from
/lib/modules/`uname -r`/modules.devname in tmpfiles.d(5) format.
Use this functionality to let systemd-tmpfilesd create the static device nodes
at boot, and drop the functionality from systemd-udevd.
As an effect of this we can move from systemd-udevd to systemd-tmpfiles-setup-dev:
* the conditional CAP_MKNOD (replaced by checking if /sys is mounted rw)
* ordering before local-fs-pre.target (see 89d09e1b5c)
During re-execution and shutdown cgroups agents might not be able
to connect to systemd's private D-Bus socket, the printed error to
the console is misleding in that case, so turn it into a warning.
During shutdown, when we try to clean up all remaining processes, the
kernel will fork new agents every time a cgroup runs empty. These
new processes cause delays in the final SIGTERM, SIGKILL logic.
Apart from that, this should also avoid that the kernel-forked binaries
cause unpredictably timed access to the filesystem which we might need to
unmount.
Lennart Poettering